User-held personalised information for routine care of people with severe mental illness(review)

Transcription

1 Cochrane Database of Systematic Reviews User-held personalised information for routine care of people with severe mental illness(review) FarrellyS,BrownGE,FlachC,BarleyE,LaugharneR,HendersonC FarrellyS,BrownGE,FlachC,BarleyE,LaugharneR,HendersonC. User-held personalised information for routine care of people with severe mental illness. Cochrane Database of Systematic Reviews 2013, Issue 10. Art. No.: CD DOI: / CD pub2. User-held personalised information for routine care of people with severe mental illness(review) Copyright 2013 The Cochrane Collaboration. Published by John Wiley& Sons, Ltd.

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY SUMMARY OF FINDINGS FOR THE MAIN COMPARISON BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure Figure Figure DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES WHAT S NEW HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST SOURCES OF SUPPORT DIFFERENCES BETWEEN PROTOCOL AND REVIEW INDEX TERMS i

3 [Intervention Review] User-held personalised information for routine care of people with severe mental illness Simone Farrelly 1, Gill E Brown 2, Clare Flach 3, Elizabeth Barley 4, Richard Laugharne 5, Claire Henderson 6 1 Health Service and Population Research Department, Institute of Psychiatry, London, UK. 2 Faculty of Health and Social Care, Edge Hill University, Ormskirk, UK. 3 Biostatistics, Health Sciences-Methodology, University of Manchester, Manchester, UK. 4 Florence Nightingale School of Nursing and Midwifery, King s College London, London, UK. 5 Mental Health Research Group, Cornwall Partnership NHS Foundation Trust and Peninsula College of Medicine and Dentistry, Exeter, UK. 6 Health Service and Population Research Department, Institute of Psychiatry, King s College London, London, UK Contact address: Simone Farrelly, Health Service and Population Research Department, Institute of Psychiatry, King s College London, De Crespigney Park, London, SE5 8AF, UK. simone.farrelly@kcl.ac.uk. Editorial group: Cochrane Schizophrenia Group. Publication status and date: New search for studies and content updated (conclusions changed), published in Issue 10, Citation: Farrelly S, Brown GE, Flach C, Barley E, Laugharne R, Henderson C. User-held personalised information for routine care of people with severe mental illness. Cochrane Database of Systematic Reviews 2013, Issue 10. Art. No.: CD DOI: / CD pub2. Background A B S T R A C T It is important to seek cost-effective methods of improving the care and outcome of those with serious mental illnesses. User-held records, where the person with the illness holds all or some personal information relating to the course and care of their illness, are now the norm in some clinical settings. Their value for those with severe mental illnesses is unknown. Objectives To evaluate the effects of personalised, accessible, user-held clinical information for people with a severe mental illness (defined as psychotic illnesses). Search methods We updated previous searches by searching the Cochrane Schizophrenia Group Trials Register in August This register is compiled by systematic searches of major databases, and handsearches of journals and conference proceedings. Selection criteria We included all relevant randomised controlled trials (RCTs) that: i. have recruited adult participants with a diagnosis of a severe mental illness (specifically psychotic illnesses and severe mood disorders such as bipolar and depression with psychotic features); and ii. compared any personalised and accessible clinical information held by the user beyond standard care to standard information routinely held such as appointment cards and generic information on diagnosis, treatment or services available. Data collection and analysis Study selection and data extraction were undertaken independently by two authors and confirmed and checked by a third. We contacted authors of trials for additional and missing data. Where possible, we calculated risk ratios (RR) and 95% confidence intervals (CI). We used a random-effects model. We assessed risk of bias for included studies and created a Summary of findings table using GRADE. 1

4 Main results Four RCTs (n = 607) of user-held records versus treatment as usual met the inclusion criteria. When the effect of user-held records on psychiatric hospital admissions was compared with treatment as usual in four studies, the pooled treatment effect showed no significant impact of the intervention and was of very low magnitude (n = 597, 4 RCTs, RR 0.99 CI 0.71 to 1.38, moderate quality evidence). Similarly, there was no significant effect of the intervention in three studies which investigated compulsory psychiatric hospital admissions (n = 507, 4 RCTs, RR 0.64 CI 0.37 to 1.10, moderate quality evidence). Other outcomes including satisfaction and mental state were investigated but pooled estimates were not obtainable due to skewed or poorly reported data, or only being investigated by one study. Two outcomes (violence and death) were not investigated by the included studies. Two important randomised studies are ongoing. Authors conclusions The evidence gap remains regarding user-held, personalised, accessible clinical information for people with psychotic illnesses for many of the outcomes of interest. However, based on moderate quality evidence, this review suggests that there is no effect of the intervention on hospital or outpatient appointment use for individuals with psychotic disorders. The number of studies is low, however, and further evidence is required to ascertain whether these results are mediated by the type of intervention, such as involvement of a clinical team or the type of information included. P L A I N L A N G U A G E S U M M A R Y Patient-held clinical information for people with psychotic illnesses User-held information is where the ill person holds personal information about their care. Such records are becoming the norm in many settings and are becoming more popular with patients. This is especially the case where the person concerned is not in hospital and receives care from more than one professional. Providing people with information about their care is thought to increase their feelings of involvement in their treatment and aims to increase people s satisfaction and participation with services, ensure early treatment and prevent hospital admission. The value of user-held personal information for those with severe mental illnesses is not known however and research evaluating the effectiveness is rare. Some research suggests that while many people decline the offer of a user-held record, the majority of those who carry their records report this to be useful. Based on a search in 2011, this review includes four trials with a total of 607 people and evaluates the effects of user-held information for people with severe mental illness. In the main, the number of relevant studies is low, with poor reporting of some outcomes. Based on moderate quality evidence, the review found that user-held information did not decrease hospital admissions, and did not decrease compulsory admissions or encourage people with severe mental illness to attend appointments (when compared to treatment as usual). Other important outcomes, such as satisfaction with care, costs and effect on mental health, were not available due to the limited quality of the four studies. There is therefore a gap in knowledge and evidence regarding user-held information for people with severe mental health problems. Further evidence is also required on the different types of user-held information (for example, if it involves the mental health team and what type of information is included in the record). Large-scale, well-conducted and well-reported studies are required to assess the effects of user-held information for people with mental illness. Two important randomised studies are currently taking place. For the present, despite a gap in evidence, user-held information is low cost and acceptable to patients, so its use is likely to grow. However, it cannot be assumed that user-held information is of benefit to people and is cost-effective without further largescale, well-conducted and well-reported trials. This plain language summary has been written by a consumer, Benjamin Gray: Rethink Mental Illness. ben.gray@rethink.org 2

5 S U M M A R Y O F F I N D I N G S F O R T H E M A I N C O M P A R I S O N [Explanation] User- held information versus Sstandard Information for routine care of people with severe mental illness Patient or population: patients with routine care of people with severe mental illness Settings: inpatient and community (UK) Intervention: user-held information versus standard information Outcomes Illustrative comparative risks*(95% CI) Relative effect (95% CI) Psychiatric hospital admission: psychiatric admission Numbers admitted Follow-up: mean 13 months Assumed risk Corresponding risk Control USER- HELD INFORM A- TION versus STAN- DARD INFORM ATION Low 1 50per1000 High 1 440per per1000 (35 to 69) 436per1000 (312 to 607) RR0.99 (0.71 to 1.38) No of Participants (studies) 597 (4 studies) Quality of the evidence (GRADE) moderate 2 Comments Psychiatric hospital admission: compulsory admission Numbers admitted Follow-up: mean 13 months Low 1 40per1000 High 1 26per1000 (15 to 44) RR0.64 (0.37 to 1.1) 507 (3 studies) moderate 3 260per per1000 (96 to 286) 3

6 Death: causes other than suicide - not reported See comment See comment Not estimable - See comment No study reported this outcome Violence - not reported See comment See comment Not estimable - See comment No study reported this outcome M ental state: psychopathology Satisfaction with health care See comment See comment Not estimable 0 (3 studies) See comment See comment Not estimable 0 (3 studies) See comment See comment Data not pooled: studies used inconsistent/ incompatible measures of psychopathology and/or data are skewed. No indication of differences between groups Data not pooled: studies used inconsistent/ incompatible measures of satisfaction and/ or data areskewed. No indication of differences between groups * The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and therelativeeffect of the intervention (and its 95%CI). CI: Confidence interval; RR: Risk ratio. GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. M oderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Lowquality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. 1 Assumed risk: the low and high risk values correspond to the lowest and highest rates of admissions in the control groups from the included studies. 4

7 2 Imprecision: rated serious - The 95% confidence interval around the pooled estimate is wide, it contains no effect, appreciable benefit and appreciable harm as defined by the GRADE system as a 25% risk decrease or increase respectively. 3 Imprecision:rated serious -The 95%confidence interval around the pooled estimate is wide,containing both no effect and an appreciable benefit as defined by the GRADEsystem as 25%risk reduction. There were also only 67 events in total. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 5

8 B A C K G R O U N D It is important to seek cost-effective methods of improving the care and outcome of those with serious mental illnesses. Userheld records, where the person with the illness holds all or some personal information relating to the course and care of their illness, are now the norm in some clinical settings (Kaelber 2008). This is particularly the case where the person concerned is not in hospital and receives care from more than one professional. User-held notes are becoming popular for obstetric care (Brown 2004; Phipps 2001) both in developed and developing countries (Shah 1993). Parent-held records are widely used in paediatric care (Bourgeois 2009; Ghossein 1998; Hampshire 2004). The management of people with chronic disabilities such as stroke or cancer can also involve user-held records (Banet 1997; Finlay 1999; Ko 2010; Williams 2001). Brief types of user-held information are available for use in emergencies such as anaphylaxis and post-splenectomy infection (Freeman 1998; Wharry 1996; Williams 1996). This type of information tends not to be comprehensive and the aims vary with setting and condition. Nevertheless, as with the full notes, the aim of these truncated personal record systems is to promote patient participation in care, increase consumer satisfaction with services, facilitate early treatment and help avoid hospital admission. Both long and short forms of user-held information are now being used for people with mental health problems (Backlar 1996; Essex 1990; Henderson 2004; Lester 2003; Pickersgill 1998; Reuler 1991; Stafford 1997; Sutherby 1999; Swanson 2006a; Wolf 1996). In certain countries legislation obligates the provision of a copy of the treatment outline and professionals contact details to those in receipt of specialist mental health services (DoH 1990). This information is not specifically designed to be something patients might wish to carry, nor to be of use in a crisis, but it may in itself have an effect, so evaluation is justified. Another form of personal information has recently emerged. Advance Statements can be defined as statements regarding service user preferences for future treatment (Henderson 2008). These are typically devised and held by the service user and include such interventions as Psychiatric Advance Directives, Joint Crisis Plans (Sutherby 1999) and Wellness Recovery Actions Plans (Cook 2009). Attempts at evaluation of the effects of user-held records for those with long term mental illness are rare. There are several audit and case series studies of the use of user-held records (Essex 1990; Pickersgill 1998; Reuler 1991; Stafford 1997; Wolf 1996). These suggest that while many people refuse the offer of a user-held record, the majority of those who carry their records report this activity to be useful (Stafford 1997). In the USA, Psychiatric Advance Directives (PADs) have been shown to be popular with service users and their carers (Backlar 2001) and some professional groups (Atkinson 2004). However, there are some concerns regarding the implementation of wishes contained in PADs during crises from both service users and clinicians (Atkinson 2004; Backlar 2001); the ability of the service user to generate a valid PAD (Amering 2007; Atkinson 2004; Elbogen 2007); and questions regarding their impact on or relationship to outcome (Backlar 2001). Additionally, clinicians often express concern regarding the possibility of service users requesting unreasonable treatments or indeed refusing treatment altogether (Atkinson 2004; Backlar 2001; Elbogen 2006; Swartz 2005; Van Dorn 2006). However, an analysis of the content of PADs developed by 106 outpatients in the US (Srebnik 2005) showed that in 95% of cases the PADs were rated as clinically useful and consistent with clinical practice standards. Another type of advance statement, the Wellness Recovery Action Plan (WRAP) is a structured self-management intervention that assists users both manage symptoms and consider broader life issues. A recent assessment of WRAPs for mental illness (Cook 2009) suggests that it may generate improvements in a range of areas including symptoms, hopefulness and feelings of empowerment. Furthermore, a case series study of the Joint Crisis Plan intervention (Sutherby 1999) in which patients develop their personalised plan for use in a crisis or possible relapse, suggests that hospital use may be reduced in those who develop such a plan. The increased input in determining future treatment was suggested to be an important factor in determining the perceived coerciveness of the treatment experience (Lidz 1995). It certainly seems plausible that those who have little knowledge of the patient or no access to their notes, such as a junior psychiatrist in a hospital emergency department, would find such information useful in making an appropriate disposition, and further that actions carried out which have been previously negotiated are less likely to require compulsory interventions. Description of the condition For the purpose of this review, we defined severe mental illness (SMI) as psychotic disorders and severe mood disorders such as bipolar disorder and depression with psychotic features. Psychosis is a collective term used to describe a range of conditions including schizophrenia, delusional disorder, and schizoaffective disorder. Individuals with psychotic disorders may experience a range of symptoms including unusual thoughts or delusions, hallucinations, and mood disturbance such as emotional flatness and withdrawal. Recent studies of schizophrenia globally suggest that about seven individuals per 1000 will be affected and individuals with schizophrenia have a two to threefold increased risk of dying compared to the general population (McGrath 2008). The economic burden associated with schizophrenia is substantial, with recent estimates suggesting it represents between 1.5% and 3% of total healthcare expenditures (Knapp 2004). Description of the intervention 6

9 User-held records may be defined as information, held by the individual with the illness, which contains all or some personal information relating to the course and care of his or her illness. User-held records may include notes made by professionals at appointments and kept by the patient, advance directives and crisis cards (personal information held for use in the event of a crisis or relapse). How the intervention might work The main aim of providing individual patients with personalised information about their illness and treatment is to improve the individual s sense of control and empowerment. Providing patients with information is thought to increase their feelings of involvement in their treatment, reducing the perceived coerciveness of the treatment experience (Lidz 1995); it may also facilitate early help-seeking. Criteria for considering studies for this review Types of studies All relevant randomised, controlled trials (RCTs). Types of participants Adults with SMI. For the purposes of the review we defined SMI as a diagnosis of a psychotic illness, including other psychoses such as bipolar disorder and depression with psychotic features. In studies where there was a mixture of diagnostic groups, only those studies where the majority of participants (that is, more than 50%) had psychotic diagnosis were included. We did not include studies where the sole diagnosis was bipolar disorder or depression. People whose main problem and primary diagnosis was one of deliberate self harm were also not the focus of this review. Types of interventions Why it is important to do this review This is an update of a Cochrane review first published in 1999 (Henderson 1999). The previous review found no studies meeting the criteria for inclusion, but found two ongoing trials (Lester 2003; Papageorgiou 2002). Since the publication of the original Cochrane review, there have been a number of published and relevant clinical trials undertaken. Additionally, current national treatment guidelines in the United Kingdom (NICE 2009) recommend involving mental health service users in decisions about their treatment and providing opportunities for service users to make advance statements and advance decisions to refuse treatment. Valid advance decisions are now legally binding under the UK s Mental Capacity Act User-held records potentially provide a cost-effective methodology for facilitating such involvement. Several of the authors of this updated review are affiliated with an ongoing trial of user-held records (CRIMSON) (Thornicroft 2010). Another randomised evaluation of user-held records for those with serious mental illnesses (Ruchlewska 2009) is known to be ongoing. O B J E C T I V E S To evaluate the effects of personalised, accessible, user-held clinical information for people with a severe mental illness (defined as psychotic illnesses). M E T H O D S 1. User-held information: any personalised and accessible clinical information held by the patient beyond standard care. This includes both user-held records (notes made by professionals at appointments and kept by the patient) and crisis cards (personal information held for use in the event of a crisis or relapse). Generic information on diagnosis, treatment or services available was excluded. 2. Standard information: any information routinely held such as appointment cards and generic information on diagnosis, treatment or services available. In certain settings standard information may include a copy of the treatment plan with contact details for the key carers (DoH 1990). Types of outcome measures Outcomes were not part of the criteria for including studies in the review. Outcomes of interest are listed below. We divided outcomes into very short term (less than three months), short term (less than six months), medium term (seven to 12 months) and long term (over one year). Primary outcomes 1. Psychiatric hospital admission 1.1 Admitted or not 1.2 Compulsory admission 7

10 1.3 Number of days spent in hospital 1.4 Discharged or not 2. Death from causes other than suicide 8. Economic costs of all care and health care 9. Other relevant measures 3. Violence 3.1 To self Non-fatal Fatal 3.2 To others Major (homicide, sex attacks, attempted or actual serious assault) Non-major (incidents requiring attendance of police or onward seclusion or special civil law admissions to a place of safety) Secondary outcomes 1. Rates of criminal charges 10. Summary of findings table We used the GRADE approach to interpret findings (Schünemann 2008) and used the GRADE profiler (GRADEPRO) to import data from RevMan 5.1 (Review Manager) to create Summary of findings tables. These tables provided outcome-specific information concerning the overall quality of evidence from each included study in the comparison, the magnitude of effect of the interventions examined, and the sum of available data on all outcomes we rated as important to patient care and decision making. We aimed to select the following main outcomes for inclusion in the Summary of findings table. Psychiatric hospital admission: general admission - medium and long term Psychiatric hospital admissions: compulsory admission - medium and long term Death: causes other than suicide Violence Mental state: psychopathology - medium and long term Satisfaction with health care - medium and long term 2. Mental state 2.1 Relapse of psychotic illness 2.2 Mental state score 3. Satisfaction with health care 3.1. Patient satisfaction 3.2. Carer satisfaction 4. Perceived coercion on hospital admission 5. Acceptability of management 5.1 As measured by loss to follow-up within the study Search methods for identification of studies Electronic searches We searched the Cochrane Schizophrenia Group Trials Register (August 2011) using the phrase: [((*consumer?h?ld* or *consumer?particip* or *client?h?ld* or *client?particip* or *user?h?ld* or *user?particip? or *patient?h?ld* or *patient particip* or (*Cris?s* AND *plan*) in the title, abstract and index terms in REFERENCE) and ((patient particip* or patient info* or *decision* or medical record* or *crisis plan*) in interventions field in STUDY)] This register is compiled by systematic searches of major databases, handsearches of journals and conference proceedings (see Group Module). For previous searches please see Appendix Compliance with treatment other than the intervention Searching other resources 7. Social functioning 7.1 Homelessness 7.2 Employment 7.3 Average change in social functioning 1. Reference lists All references of included articles were searched for further relevant trials. 8

11 Data collection and analysis Methods used in data collection and analysis for this update are below, for previous methods please see Appendix 2. Selection of studies SF and GB independently inspected all citations from the searches and identified relevant abstracts. Where disputes arose, the full report was acquired for more detailed scrutiny and was independently assessed. All included citations were then discussed with CH. Where it was not possible to resolve a disagreement by discussion, we contacted the authors of the study for clarification. If citations met the inclusion criteria, we obtained full reports of the papers for more detailed inspection. Full reports were then inspected by SF and GB, and discussed with CH. Data extraction and management 1. Extraction Review author SF extracted data from all included studies. In addition, to ensure reliability, GB independently extracted data from all studies. Any disagreement was discussed with CF (statistician) and the decisions documented and, if necessary, authors of studies were contacted for clarification. CH provided advice and final decisions were documented. We attempted to contact authors through an open-ended request in order to obtain missing information or for clarification whenever necessary. If studies were multi-centre, where possible, we extracted data relevant to each component centre separately. 2. Management 2.1 Forms We extracted data onto standard, simple forms. 2.2 Scale-derived data We included continuous data from rating scales only if: a. the psychometric properties of the measuring instrument have been described in a peer-reviewed journal (Marshall 2000); and b. the measuring instrument has not been written or modified by one of the trialists for that particular trial. Ideally the measuring instrument should either be: i. a self-report, or ii. completed by an independent rater or relative (for example, not the therapist). We acknowledge that this is often not reported clearly and noted if this was the case or not in the Description of studies section. 2.3 Endpoint versus change data There are advantages of both endpoint and change data. Change data can remove a component of between-person variability from the analysis. On the other hand the calculation of change needs two assessments (baseline and endpoint), which can be problematic in unstable and difficult to measure conditions such as schizophrenia. We decided to primarily use endpoint data, and only use change data if the former were not available. If we had found both, we would have combined endpoint and change data in the analysis and used mean differences (MD) rather than standardised mean differences throughout (Higgins 2011). 2.4 Skewed data Many of our targeted outcomes tend to be skewed, for example, duration of admission and duration of compulsory admission. For this reason the authors of included studies have provided nonparametric summary measures and tests, that is, the median and range, and tested with the Mann-Whitney U test. Since it is not possible to combine non-parametric data in a meta-analysis, and there were so few studies that met the inclusion criteria and provided these data, we report these data in Other data tables. 2.5 Common measure To facilitate comparison between trials, we converted variables reported in different metrics, such as days in hospital (mean days per year, per week or per month), to a common metric (for example, mean days per month). 2.6 Conversion of continuous to binary Where possible we tried to maintain continuous measures to avoid losing information and power in the analyses. 2.7 Direction of graphs Where possible, we entered data in such a way that the area to the left of the line of no effect indicated a favourable outcome for user-held records. Where keeping to this made it impossible to avoid outcome titles with clumsy double-negatives we reported data where the left of the line indicates an unfavourable outcome. This was noted in the relevant graphs. Assessment of risk of bias in included studies SF and GB worked independently to assess risk of bias by using the criteria described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) to assess trial quality. This set of criteria is based on evidence of associations between overestimation of effect and high risk of bias of the article, such as sequence generation, allocation concealment, blinding, incomplete outcome data and selective reporting. 9

12 If the raters disagreed, the final rating was made by consensus, with advice from CH. Where inadequate details of randomisation and other characteristics of trials were provided, the authors of the studies were contacted in order to obtain further information. Non-concurrence in quality assessment was reported, but if disputes arose as to which category a trial was to be allocated, again, resolution was be made by discussion. Measures of treatment effect 1. Binary data For binary outcomes we calculated a standard estimation of the risk ratio (RR) and its 95% confidence interval (CI). It has been shown that RR is more intuitive (Boissel 1999) than odds ratios and that odds ratios tend to be interpreted as RR by clinicians (Deeks 2000). For statistically significant results we had planned to calculate the number needed to treat to provide benefit/to induce harm statistic (NNTB/H), and its 95% confidence interval (CI) using Visual Rx ( taking account of the event rate in the control group. This, however, has been superseded by the Summary of findings for the main comparison and calculations therein. 2. Continuous data We had too few studies to combine continuous outcomes. Furthermore, outcomes common to more than one study were assessed by different measures therefore precluding any combined estimate of effect. However, if we had found sufficient studies we would have estimated the mean difference (MD) between groups. Data from included studies were often skewed, in these circumstances we have presented the data as reported by the original authors in Other data tables. Unit of analysis issues 1. Cluster trials Studies increasingly employ cluster randomisation (such as randomisation by clinician or practice) but analysis and pooling of clustered data poses problems. Firstly, authors often fail to account for intra-class correlation in clustered studies, leading to a unit of analysis error (Divine 1992) whereby P values are spuriously low, confidence intervals unduly narrow and statistical significance overestimated. This causes type I errors (Bland 1997; Gulliford 1999). For relevant cluster trials that did not account for clustering in the primary studies (Warner 2000), we have presented data in a table marked with a (*) symbol to indicate the presence of a probable unit of analysis error. We have sought statistical advice and were advised that the binary data as presented in a report should be divided by a design effect. This is calculated using the mean number of participants per cluster (m) and the intra-class correlation coefficient (ICC) (design effect = 1 + (m - 1)*ICC) (Donner 2002). If the ICC was not reported it was assumed to be 0.1 (Ukoumunne 1999). Clustering was appropriately incorporated into the analysis of one of the cluster studies (Lester 2003) but not in the second (Warner 2000). Therefore we have reported the original data as if from a non-cluster randomised study (Analysis 1.1) and, in addition, have provided a cluster-adjusted outcome where frequencies have been divided by the design effect (Analysis 1.8). 2. Cross-over trials A major concern of cross-over trials is the carry-over effect. It occurs if an effect (for example, pharmacological, physiological or psychological) of the treatment in the first phase is carried over to the second phase. As a consequence, on entry to the second phase the participants can differ systematically from their initial state despite a wash-out phase. For the same reason cross-over trials are not appropriate if the condition of interest is unstable (Elbourne 2002). We did not find any cross-over trials in our search. 3. Studies with multiple treatment groups We did not find any multiple treatment group studies, but we would have presented the additional treatment arms in comparisons. If data were binary we would have added and combined the data within the two-by-two table. If data were continuous we would have combined data following the formula in section (Combining groups) of the Cochrane Handbook for Systematic Reviews of Interventions. Where the additional treatment arms were not relevant, we would not have reproduced these data. Dealing with missing data 1. Overall loss of credibility At some degree of loss of follow-up data must lose credibility (Xia 2009). We chose that, for any particular outcome, should more than 50% of data be unaccounted for we would not reproduce these data or use them within analyses. If, however, more than 50% of those in one arm of a study were lost, but the total loss was less than 50%, we would have marked such data with (*) to indicate that such a result may well be prone to bias. However, this did not occur for any of the included studies. 2. Binary The included studies had low attrition on the primary outcomes. However, had we found higher attrition rates we would have followed the following procedure. Where attrition for a binary outcome was between 0% and 50% and where these data were not 10

13 clearly described, we would have presented data on a once randomised always analyse basis (an intention-to-treat analysis). We would have assumed that data were missing at random, and that those leaving the study early would be similar to those remaining and have the same rates of negative outcome as those who completed the study. 3. Statistical heterogeneity 3.1 Visual inspection We visually inspected graphs to investigate the possibility of statistical heterogeneity. 3. Continuous 3.1 Attrition If attrition for a continuous outcome had been between 0% and 50% and completer-only data were reported, we would have reproduced these. However, this was not the case. 3.2 Standard deviations The continuous outcome measures in our studies were skewed and authors reported the median as a summary measure rather than the mean. In the case of skewed data the standard deviation is not a useful measure and so we did not seek to obtain it. Had the mean values been reported but not the standard deviations we would have first tried to obtain the missing values from the authors. If not available, we would have calculated the standard deviation, where possible, from the data provided, that is, from confidence intervals and statistical tests. 3.3 Last observation carried forward We did not find any studies that used the last observation carried forward (LOCF) for missing data. If this had been used in a trial for less than 50% of the data, we would have reproduced these data and indicated that they were the product of LOCF assumptions. 3.2 Employing the I 2 statistic Heterogeneity between studies was investigated by considering the I 2 statistic method alongside the Chi 2 P value. The I 2 provides an estimate of the percentage of total variance due to study heterogeneity (beyond inconsistency thought to be due to chance) (Higgins 2003). The importance of the observed value of I 2 depends on: i. magnitude and direction of effects, and ii. strength of evidence for heterogeneity (for example, P value from Chi 2 test, or a confidence interval for I 2 ). An I 2 estimate greater than or equal to around 50%, accompanied by a statistically significant Chi 2 statistic, was interpreted as evidence of substantial levels of heterogeneity (Higgins 2011). When substantial levels of heterogeneity were found in the primary outcome, we explored reasons for heterogeneity (Subgroup analysis and investigation of heterogeneity). Assessment of reporting biases Reporting biases arise when the dissemination of research findings is influenced by the nature and direction of results (Egger 1997). These are described in Section 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We are aware that funnel plots may be useful in investigating reporting biases but are of limited power to detect small-study effects. As we had less than 10 studies, we did not use funnel plots. Assessment of heterogeneity 1. Clinical heterogeneity We considered all included studies initially, without seeing comparison data, to judge clinical heterogeneity. We simply inspected all studies for clearly outlying people or situations which we had not predicted would arise. When such situations or participant groups arose, these were discussed. 2. Methodological heterogeneity We considered all included studies initially, without seeing comparison data, to judge methodological heterogeneity. We simply inspected all studies for clearly outlying methods which we had not predicted would arise. When such methodological outliers arose these were fully discussed. Data synthesis We understand that there is no closed argument for preference for use of fixed-effect or random-effects models. The random-effects method incorporates an assumption that the different studies are estimating different, yet related, intervention effects. This often seems to be true to us and the random-effects model takes into account differences between studies even if there is no statistically significant heterogeneity. There is, however, a disadvantage to the random-effects model. It adds an extra weight to all studies, making the weights more evenly spread and giving small studies relatively more weight compared to the fixed-effect model analysis. The small studies are often the most biased ones and, depending on the direction of effect, these studies can either inflate or deflate the effect size. We chose the random-effects model for all analyses as the studies were all of similar size. The reader is, however, able to choose to inspect the data using the fixed-effect method. 11

14 Subgroup analysis and investigation of heterogeneity 1. Subgroup analyses - only primary outcomes We proposed to undertake this review and provide an overview of the effects of user-held records for people with schizophrenia, and therefore we did not perform any subgroup analyses. 2. Investigation of heterogeneity If inconsistency was high, this was reported. First we investigated whether the data had been entered correctly. Second, if data were correct, the graph was visually inspected and studies outside the company of the rest were successively removed to see if homogeneity was restored. For this review we decided that should this occur with data contributing to the summary finding of no more than around 10% of the total weighting, the data were presented. If not, the data were not pooled and issues were discussed. We know of no supporting research for this 10% cut off but are investigating use of prediction intervals as an alternative to this unsatisfactory state. When unanticipated clinical or methodological heterogeneity was obvious, we simply stated that hypotheses regarding these for future reviews or versions of this review would be generated. We did not undertake analyses relating to these. 3. Risk of bias We analysed the effects of excluding trials that were judged to be at high risk of bias across one or more of the domains of randomisation (implied as randomised with no further details available), allocation concealment, blinding and outcome reporting for the meta-analysis of the primary outcome. If the exclusion of trials at high risk of bias did not substantially alter the direction of effect or the precision of the effect estimates, then data from these trials were included in the analysis. 4. Imputed values We also undertook a sensitivity analysis to assess the effects of including data from trials where we used imputed values for ICC in calculating the design effect in cluster randomised trials. If substantial differences were noted in the direction or precision of effect estimates in any of the sensitivity analyses listed above, we did not pool data from the excluded trials with the other trials contributing to the outcome but presented them separately. 5. Fixed and random effects We assessed the differences between using a fixed-effect model and a random-effects model on our primary outcomes of interest. Sensitivity analysis 1. Implication of randomisation We planned to include trials in a sensitivity analysis if they were described in some way as to imply randomisation. For the primary outcomes we included these studies and if there was no substantive difference when the implied randomised studies were added to those with a better description of randomisation, then all data were employed from these studies. 2. Assumptions for lost binary data There was low attrition in the included studies and therefore it was not necessary to perform a sensitivity analysis. R E S U L T S Description of studies Results of the search Electronic searches identified 237 references. Seventy-five potentially relevant records were obtained and scrutinised, and 70 of these reports did not meet the inclusion criteria (see Characteristics of excluded studies). Two references were protocols for ongoing trials for which no data were available at the time of review (Figure 1). 12

15 Figure 1. Study flow diagram. 13

16 Included studies Four studies met our inclusion criteria (Henderson 2004; Lester 2003; Papageorgiou 2002; Warner 2000) and are summarised in the Characteristics of included studies table. One study met our inclusion criteria, however we were unable to include it as the required summary statistics and results were not provided in the report. We have classified this study (Swanson 2006) as awaiting classification in the event we should obtain the relevant results from the study authors in the future had 63% with psychosis and 28% with depression and bipolar. Warner 2000 had 42% schizophrenia and 12% bipolar (the remaining diagnoses included 22% depression and 14% other ) and thereby reached our inclusion criterion (at least 50%). In terms of gender profile, there were 338 male participants and 269 female participants across the studies, with the studies having a majority male sample (range 53% to 68%). The average age in the studies was very similar, with inclusion criteria ranging from 18 to 66 years, and a mean age of 39 years. 1. Study design All studies were randomised controlled trials. Warner 2000 and Lester 2003 used cluster randomisation and the remaining trials used individual level randomisation. Both cluster trials acknowledged and investigated clustering effects within their reported data. The Lester 2003 study reported that the random-effects model did not alter the results of their analysis and so reported the original figures, which we use in our report. The Warner 2000 study did not specifically report a cluster-adjusted analysis, but provided a design effect based on the intra-class correlation coefficient of We have reported the results with the original data and have additionally provided a cluster-adjusted result where the Warner 2000 frequencies have been divided by the design effect. 2. Study sizes The total number of participants included in the four studies was N = 607. The average among the studies was n = 152 and the range was 90 to Study setting All studies were conducted in the United Kingdom. Warner 2000 and Lester 2003 were conducted in a shared care setting, that is, care that was jointly provided by primary and secondary care. Lester 2003 was conducted between six community mental health teams in Birmingham (UK); Papageorgiou 2002 was delivered in an inpatient setting, whereas all other trials involved participants who were treated in community settings. Participants in Henderson 2004 were from eight community mental health teams in South London and Kent (UK). 4. Participants With the exception of Lester 2003, which included only those people with a diagnosis of schizophrenia, all studies had a mixture of diagnoses. Henderson 2004 recruited individuals with psychosis and bipolar disorders (proportions not reported). Papageorgiou 5. Interventions Two of the studies (Henderson 2004; Papageorgiou 2002) tested a form of advance statement (Henderson 2008); that is, personal information and treatment preferences held for use in the event of a crisis or relapse. In these two studies, the user developed the user-held record with the help of a researcher or facilitator, and in the Henderson 2004 study the clinical team involved in delivering routine care was also involved. The content of these interventions was similar and broadly included the users views on relapse indicators, wishes for future treatment, contact details for themselves, their clinicians and family and carers. They may also have included an indication of treatments that they did not want, that is, a refusal of medication. The other two studies (Lester 2003; Warner 2000) tested a form of medical record which included aspects such as contact details, clinical notes, future appointments and medication. Lester 2003 also included an indication of relapse indicators and a diary section. For Lester 2003 and Warner 2000, these records were given to the user after randomisation. All studies compared the active intervention to treatment as usual in primary or secondary care. In addition, Henderson 2004 provided control participants with information leaflets regarding local services, treatments, and relevant legislation and policies (Henderson 2004). 6. Outcomes (rating scales) A range of outcomes were investigated in the included studies. However, only four outcomes were investigated by more than one study: psychiatric hospital admissions; psychopathology; satisfaction; and outpatient attendance. All outcomes investigated by the studies are listed below. 6.1 Psychiatric admissions Each included study compared the effect of user-held records to treatment as usual on the rate and length of psychiatric admission. 14

17 6.2 Compulsory psychiatric admissions Henderson 2004, Lester 2003 and Papageorgiou 2002 also examined the rates of compulsory psychiatric hospital admissions under a section of the UK s Mental Health Act. Henderson 2004 and Papageorgiou 2002 additionally looked at number of days spent under a section of the Mental Health Act Client Satisfaction Questionnaire (CSQ) (Larsen 1979) Warner 2000 used the Client Satisfaction Questionnaire (CSQ), which is an eight-item self-report measure covering: quality of service; needs addressed; amount of assistance; effect of assistance; recommendable; and overall satisfaction. Each item is rated on a one to four scale, with a lower score indicating a worse outcome. 6.3 Psychopathology Psychopathology was examined by three of the four studies (Lester 2003; Papageorgiou 2002; Warner 2000) Brief Psychiatric Rating Scale (BPRS) (Overall 1962) Warner 2000 used the BPRS, which is a rating scale designed to assess psychopathology across 16 dimensions: somatic concern; anxiety; emotional withdrawal; conceptual disorganisation; guilt feelings; tension; mannerisms and posturing; grandiosity; depressive mood; hostility; suspiciousness; hallucinatory behaviour; motor retardation; uncooperativeness; unusual thought content; and blunted affect. The BPRS is usually rated by an experienced researcher or clinician, with each item rated on a seven-point scale, varying from 0 = not present to 7 = extremely severe, with high scores indicating more severe symptoms Behavior and Symptom Identification Scale (BASIS-32) (Eisen 1994) Papageorgiou 2002 and Warner 2000 used the BASIS-32 scale, which is a self-report measure completed by patients. The BASIS- 32 has five subscales: relation to self and others; daily living and role functioning; depression and anxiety; impulsive and addictive behaviour; and psychosis. Thirty-two items are rated between 0 (= no problem) and 4 (= severe problem) Krawiecka and Goldberg (K & G) scale (Krawiecka 1977) Additionally, Lester 2003 used the K & G scale, which is a fivepoint rating scale assessing psychopathology across the following domains: depressed; anxious; coherently expressed delusions; hallucinations; incoherence and irrelevance of speech; poverty of speech; flattened incongruous affect; psychomotor agitation; and side effects of medication (tremor, rigidity, dystonic reactions, akathisia, difficulties with vision, other). A score of 0 represents the absence of an item, where a score of 4 may indicate severe psychopathology. 6.4 Satisfaction Verona Service Satisfaction Scale (VSSS) (Ruggieri 1993) Three studies examined the effect of the intervention on patient satisfaction. Lester 2003 and Papageorgiou 2002 used the Verona Service Satisfaction Scale (VSSS), however Papageorgiou 2002 used an adapted brief version. The full VSSS-54 contains 54 items rating questions over seven dimensions: overall aspects; professionals skills and behaviour; information; access; efficacy; types of intervention; and relative s involvement. This is a five-point Likert scale, with a score of 1 = terrible and 5 = excellent. 6.5 Outpatient visits Lester 2003 and Warner 2000 both examined outpatient visits in terms of attendance. 6.6 Economic costs The Henderson 2004 study published a secondary economic evaluation paper comparing the costs of service use between those who received the intervention and the control group over the 15- month study period. They used a modified version of the client service receipt inventory to collect information about use of services. Hospital use was collected from patient medical records and the Mental Health Act office. 6.7 Other outcomes Several other outcomes were investigated by individual studies including: self efficacy (no summary statistics so unable to include) (Papageorgiou 2002); and use of home treatment and non-mental health referrals (Lester 2003). Excluded studies Of the 75 potentially relevant references identified in the updated search, 17 were duplicate studies or secondary references, and 49 were excluded as they did not meet the inclusion criteria described in the Types of studies, Types of participants, and Types of interventions. The most common reason for exclusion was that the intervention did not involve a user-held information component. Four further references were excluded after contacting the authors (n = 3) or having materials translated (n = 1). Hamann 2006 was a trial of shared decision making compared with standard treatment for 113 individuals with psychotic disorders treated in an inpatient setting in Germany. The intervention involved patients working through a decision aid booklet and noting their preferences for care, and using that as a prompt for discussion with their treating doctor; however, the patients did not hold the a record after this consultation. Similarly, Van Os 2004 examined the effect of patients completing a checklist (2-COM) to identify care needs prior to routine consultations with their doctor, however the patients did not hold a record of the consultation or treatment plan. Likewise, Borell 1995 examined a psycho-education initiative that involved some personalised information, but the patients did not hold the records after the sessions. Finally, Swanson 2008 was excluded as it was a secondary paper from the Swanson 2006 study 15

18 that examined coercive events within the intervention group who received a Facilitated Psychiatric Advance Directive, that is, there was no comparison group. Previously excluded studies are discussed below and reasons for exclusion are listed in the Characteristics of excluded studies table. Stafford 1997 was an audit of 45 people with long term mental illness who agreed to carry a small pocket-sized file containing contact details, early warning signs for relapse, and notes made by professionals. Level of use was recorded by checking entries in the professionals notes against those in the record, and the holders were surveyed on their views of the usefulness of the record as a whole and of its separate components. There was no control group. Two randomised studies concerned the use of green cards to facilitate access to services for patients with a primary diagnosis of deliberate self harm (Cotgrove 1995; Morgan 1993) rather than one of a psychotic illness. Six other studies were already known of by the authors but were not identified by the search strategy. The aims of three (Essex 1990; Pickersgill 1998; Wolf 1996) were to improve shared care between professional primary care givers and specialist mental health services as well as to empower users of mental health services (Essex 1990; Reuler 1991; Wolf 1996). Pickersgill 1998 primarily focused on patient empowerment along with developing partnerships with users and carers. Each of these uncontrolled studies described the development and introduction of a personal record and then surveyed the individuals views on its usefulness or level of use. (Sutherby 1999) described the use of crisis cards and joint crisis plans for people with severe mental illnesses who were in receipt of specialist care. Joint crisis plans were drawn up once the contents had been agreed by both the team and patient at a meeting between the patient, an advocate if required, team members and a facilitator. Where the team could not agree with the patient s wishes, these were to have been written on a crisis card instead. Participants were then followed up and asked their views on the plans and whether they had affected how they felt about their illness and treatment. There was no control group. The last case series (Backlar 1996) was not an intervention study but a postal survey of members of organisations for relatives of mentally ill people, to assess the level of use of advance directives for mental health treatment. Ongoing Two ongoing studies were found (Ruchlewska 2009; Thornicroft 2010) and are summarised in the Characteristics of ongoing studies table. Both trials were due to be completed at the end of 2011 and were expected to report results in Awaiting assessment Swanson 2006 did not report any summary statistics of potentially relevant outcomes. We were unable to obtain further data from the authors. Risk of bias in included studies For summary and graphical representation of the risk of bias please see Figure 2 and Figure 3. Figure 2. Methodological quality graph: review authors judgements about each methodological quality item presented as percentages across all included studies. 16